[U-Boot-Users] Running test code from RAM while having GOT pointing to flash

[U-Boot-Users] Running test code from RAM while having GOT pointing to flash

[atul.sabharwal at exgate.tek.com]

I am a newbie at PowerPC assembly.  In my last patch, I submitted a series
of test as part of my work in bringing up the memory controller.

My code seems to be crashing in the relocate_code routine while switching
from flash to RAM.  There was an address_probe fixup which was not enabled
And I enabled it. So, the relocation table (GOT) gets copied successfully

[U-Boot-Users] Running test code from RAM while having GOT pointing to flash

[Wolfgang Denk]

In message <4A062D477D842B4C8FC48EA5AF2D41F20152806A@us-bv-m23.global.tektronix.net> you wrote:
> My code seems to be crashing in the relocate_code routine while switching
> from flash to RAM.  There was an address_probe fixup which was not enabled
> And I enabled it. So, the relocation table (GOT) gets copied successfully
> From flash to RAM but the code dies seomewhere between coping the fixups

I think you misunderstand the function of the GOT.

> And the in_ram label.  Somehow, on my BDI, it sees an exception and goes
> to the reset vector i.e 0x100.

See the FAQ at http://www.denx.de/wiki/view/DULG/UBootCrashAfterRelocation

> My approach is to avoid the GOT stuff and write a test routine which is

The "GOT stuff" is known to work just fine. I suggest  you  focus  on
the  remaining  stuff, i. e. your SDRAM init sequence. Again, see the
FAQ.

Best regards,

Wolfgang Denk

-- 
Software Engineering:  Embedded and Realtime Systems,  Embedded Linux
Phone: (+49)-8142-66989-10 Fax: (+49)-8142-66989-80 Email: wd at denx.de
I often quote myself; it adds spice to my conversation.  - G. B. Shaw

[U-Boot-Users] Running test code from RAM while having GOT pointing to flash

[Jerry Van Baren]

atul.sabharwal at exgate.tek.com wrote:
> I am a newbie at PowerPC assembly.  In my last patch, I submitted a series
> of test as part of my work in bringing up the memory controller.

[snip]

> My first cut at it is here but it is complaining about GOT. Not enough
> experience with ld...   Help/tips ??
> 
> 
>    __asm__ __volatile__(		   \
> 		      "	lis r10, 0x0 at h\n"		   \
> 		      "	eieio\n"		           \
> 		      " lis r3, address at l\n"		   \
> 		      " ori r3, r3, address at h\n"           \
> 		      " lis r4, fillpattern at l\n"           \
> 		      " ori r4, r4, fillpattern at h\n"       \
> 		      " lis r5, bytecount at l\n"             \
> 		      " ori r5, r5, bytecount at h\n"         \
> 		      " b  0x0\n"                         \
> 		      );
> 
> Thanks,
> 
> Atul

This has nothing to do with your current problems, but if you try to run 
the above code, it won't work.  You are doing a "load shifted" of the 
_lower_ half of the word and then oring in the _upper_ half of the word. 
    You need to switch your "@l" (low half) and "@h" (high half).  What 
you have will give you a halfword-swapped value:

Want: 01234567
Oops: 45670123

gvb